Multimodal large language models (MLLMs) are vulnerable to visual jailbreak attacks that bypass safety safeguards. Existing methods typically optimize isolated adversarial examples, limiting attack diversity, transferability, and generalization. We propose Jailbreak Connectivity (JC), a framework based on the observation that effective adversarial inputs can be connected through low-loss regions in image space. JC optimizes continuous adversarial paths: path sampling yields diverse jailbreak images, lightweight surrogate classifiers guide cross-model transfer, and a distribution-level objective enables universal attacks across harmful queries. On MM-SafetyBench, JC achieves average attack success rates of 84.01%, 85.77%, and 97.95% on Qwen3-VL-8B-Instruct, InternVL3.5-8B, and Gemma-3-12B-IT, respectively, exceeding the strongest baselines by up to 68.36 percentage points. In black-box evaluations, JC improves attack success rates over FigStep by 24.69 and 27.34 percentage points on GPT-5.6 and Gemini 3.1 Pro, respectively. These findings reveal structured vulnerabilities in MLLM safety alignment and inform more comprehensive security evaluation.
LLM agents in interactive entertainment and social applications, including companions, role-play characters, and persona clones, need decisions grounded in personality and past experiences. We introduce HEART-Bench to evaluate memory-grounded persona-consistent response selection. The benchmark combines 11 Big Five-anchored synthetic characters, each with 1,000 autobiographical memories, and 64 DIAMONDS-guided decision scenarios into 673 multiple-choice questions. Domain experts independently compare candidate responses, with a senior expert resolving disagreements and revising responses when needed; other characters' adjudicated responses to the same scenario serve as distractors. In the main evaluation, models receive character IDs and occupations, scenarios, and options, with retrieved memories added in memory-augmented settings. Detailed personality-and-value profiles are withheld. We evaluate 16 LLMs under No-Retrieval, Naive-RAG, Mem0, and PersonaDB. Accuracy without memories stays near the 25% chance level, and the best memory-augmented configuration reaches 64.9%. On a matched 150-item subset, human annotators reach 80.8%, compared with 64.0% for the best model. Memory-content ablations show the importance of retrieving experiences that match both the target character and the decision context.
Zhaoze Wang, Changxu Zhang, Tai Fei, Christopher Grimm, Yi Jin, Claas Tebruegge, Markus Gardill
Modern simulators can generate synthetic camera and LiDAR data with accurate annotations for downstream tasks. In contrast, automotive radar simulation remains less mature because it requires not only a 3D engine to model dynamic objects in the scenes but also electromagnetic material properties, antenna configurations, wave propagation mechanisms, and coherent signal synthesis. Existing frameworks are either proprietary or overly simplified, limiting the scalability of radar raw measurements. This paper presents AuRaSim, a ray-tracing-based automotive frequency-modulated continuous wave (FMCW) radar simulation framework for dynamic scenarios. Rather than implementing all components from scratch, AuRaSim builds on Blender for efficient dynamic-scene construction and Sionna RT for graphics processing unit (GPU) accelerated electromagnetic ray tracing, while introducing radar-specific adaptations for path-resolved propagation with complex coefficients, motion-aware Doppler modeling, and subsequently coherent multi-channel analog-to-digital converter (ADC) cube synthesis based on waveform configurations. Quantitative comparison between simulated and measured range-Doppler maps in digital-twin scenarios achieves a root-mean-square error (RMSE) of 0.147 and a peak signal-to-noise ratio (PSNR) of 16.93 dB, demonstrating the accuracy and fidelity of the synthetic signal. AuRaSim therefore provides a complete and extensible platform for reproducible automotive radar simulation and research. Project homepage: https://wangzhaoze.github.io/aurasim.
Remote sensing (RS) image-text retrieval plays a critical role in understanding massive RS imagery. However, the dense multi-object distribution and complex backgrounds in RS imagery make it difficult to simultaneously achieve fine-grained cross-modal alignment and efficient retrieval. Existing methods either rely on complex cross-modal interactions that lead to low retrieval efficiency, or depend on large-scale vision-language adaptation, which requires massive data and computational resources. To address these issues, we develop a fast-then-fine (FTF) retrieval framework centered on interaction-free multi-granular recall and parameter-free text-guided cross-modal reranking with the Balanced Text-guided Interaction Block (BTIB). Specifically, interaction-free multi-granular recall pre-computes gallery embeddings and matches them with the corresponding query embedding without full-gallery token-level text-image interaction, while text-guided cross-modal reranking with BTIB refines local semantic correspondence within the top-K candidate set. Furthermore, an inter- and intra-modal loss is designed to jointly optimize cross-modal alignment across multi-granular representations. Extensive experiments on public benchmarks demonstrate that the FTF achieves competitive retrieval accuracy while significantly improving retrieval efficiency compared with existing methods.
Zhangkai Wu, Xu Zhang, Zhihong Cui, Yiqi Wang, Yi Zhang
Tabular foundation models (TFMs) are being applied to Web data streams such as network intrusion detection, where they adapt by updating the labelled rows of their context without retraining. When a past concept returns after leaving the context, a TFM faces a recall-or-relearn decision between restoring stored rows of the concept and relearning it from new labels. We propose MICE (Mixture of In-Context Experts), a training-free learner that cuts the stream into segments at detected changes. It merges a closed segment into a stored context when the TFM's estimated additional error on the segment is within a tolerance, under a capacity limit on stored contexts. A fast level weights the stored contexts and first-in-first-out (FIFO) windows by the log-loss of their predictions on the last labelled batch. We show that, with pure stored contexts and a margin condition, this fast level gains over relearning, in expected accuracy, the normalised positive area between the stored context's accuracy and the TFM's learning curve, up to a first-batch term. A slow level mixes the fast level's prediction with that of the FIFO window of the last labelled rows and bounds the discounted half-Brier regret of MICE against this window at every time. Learning curves measured beforehand predict the ordering of the observed recall gains. MICE outperforms reset, which empties the context at detected drifts, on grid streams with random boundary phases. Merging keeps fewer experts than pools that store every segment, and with fixed segmentation a time-matched snapshot pool matches MICE on the shifted grid. On real streams, MICE has the highest mean accuracy of all policies we compare.
Yitao Xiao, Xu Zhang, Ye Cheng, Shaoyong Guo, Yan Sun, Ziyue Zhu, Sitao Xie, Guo Tong 等 9 位作者
Serving web queries at scale with multi-agent systems requires jointly configuring large language models, collaboration modes, agent counts, interaction rounds, and skills. However, the resulting configuration space grows combinatorially, while sparse execution rewards provide limited fine-grained supervision for interdependent planning decisions. We propose EvoRouter, a self-evolving multi-agent routing framework driven by insight-guided on-policy distillation. A dedicated expert periodically synthesizes performance-cost feedback from configuration trajectories across tasks into reusable conditional routing insights about capability requirements, returns from resource scaling, and interactions among configuration choices. A privileged teacher combines these evolving insights with current execution feedback to provide dense token-level supervision for the router's own planning trajectories through on-policy distillation. Cost-aware reinforcement learning jointly optimizes task performance and computational cost. Together, these training signals internalize execution insights into the routing policy. The updated policy generates new trajectories that support subsequent insight extraction, forming an execution-driven self-improvement loop. We evaluate EvoRouter on 19 in-domain datasets spanning five domains and two out-of-domain benchmarks. EvoRouter achieves an average in-domain accuracy of 68.69%, outperforming the strongest routing baseline by 7.23 percentage points while reducing inference compute cost by 50.95%. Our code is open-sourced at https://github.com/CenYangtze/EvoRouter.
Navigational search aims to find a single target page among hundreds of millions of web pages. Click tables work well for queries seen before but cannot generalise to new formulations, while embedding retrievers generalise across content but often rank navigational targets poorly. We introduce Octen-WebGR, a generative retrieval framework that stores web pages in the parameters of a language model: given a query, the model generates a two-token semantic identifier constrained to a table of real URLs. We develop Octen-WebGR-4B, which indexes 201 million web pages, and evaluate it on real navigational traffic from the month after training. Generative retrieval is most effective on new ways of asking for familiar pages: for unseen queries targeting pages with substantial interaction history, it ranks the correct page first 30.9% of the time, against 17.2% for a dense retriever trained on the same clicks (21.4% with a click prior). Embedding retrieval keeps an advantage on rarely visited pages, so the two approaches generalise on complementary sides, the query and the page. Over all navigational traffic, Octen-WebGR reaches 59.6% top-1 accuracy, against 46.5% for a click-enhanced dense retriever and up to 37.5% for frontier LLMs answering from memory, which cost at least 40 times as much per query. The 4B model runs on a single GPU with 10 GB of memory, compared with 412 GB for exact dense retrieval. Scaling experiments with up to 32 million URLs show that, when every query-page pair is seen equally often, the model size needed grows sublinearly with index size, and that retrieval is constrained more by page exposure than by model capacity. We release Octen-WebGR-4B and Octen-WebGR-0.6B with their identifier tables and inference code.
Web agents reuse skills whose stored code, selected version, and loaded implementation can diverge across tasks and updates. SkillLifeBench links these states to task outcomes in 388 paired scenarios spanning implementation, interface, runtime, composition, and version changes. We evaluate 14 models on two paired decision tracks: 89 runtime tasks and 60 autonomous lifecycle tasks spanning 12 templates derived from introduced cases. Lifecycle tasks allow agents to inspect, select, and reconfigure skills while completing verifiable user requests. Joint success requires task completion without the registered attack goal. GPT-5.6-Sol and Sonnet 4.6 both achieve 82.0% on runtime tasks, but reach 96.7% and 56.7% on lifecycle tasks. Across lifecycle evaluations, challenged task success is 77.5%, while joint success is 65.2%; 103 runs complete the task and the attack goal. Adding execution provenance in a matched recovery study raises utility joint success from 49.35% with logs and metadata to 65.58%. A deterministic execution-consistency rule reaches 64.94% with higher benign utility, while full restoration reaches 69.48%. Lifecycle tasks distinguish models with similar runtime scores, and execution evidence improves recovery.
Fan Xu, Xinyu Hu, Zhenghan Yu, Li Lin, Xu Zhang, Yang Zhang, Wei Zhou, Jinjie Gu 等 9 位作者
The increasing reliance on natural language generation (NLG) models, particularly large language models, has raised concerns about the reliability and accuracy of their outputs. A key challenge is hallucination, where models produce plausible but incorrect information. As a result, hallucination detection has become a critical task. In this work, we introduce a comprehensive hallucination taxonomy with 11 categories across various NLG tasks and propose the HAllucination Detection (HAD) models, which integrate hallucination detection, span-level identification, and correction into a single inference process. Trained on an elaborate synthetic dataset of about 90K samples, our HAD models are versatile and can be applied to various NLG tasks. We also carefully annotate a test set for hallucination detection, called HADTest, which contains 2,248 samples. Evaluations on in-domain and out-of-domain test sets show that our HAD models generally outperform the existing baselines, achieving state-of-the-art results on HaluEval, FactCHD, and FaithBench, confirming their robustness and versatility.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 6: Industry Track) · 2026 · 正式发表
As large language models (LLMs) demonstrate remarkable capabilities across a wide range of tasks, ensuring the safety of their outputs is increasingly critical. To mitigate the risk of policy-violating responses, numerous guardrail models have been developed for harmful-content detection. While effective on short outputs, existing guardrails degrade on long-form responses, reflecting limited semantic understanding and weak robustness to contextual noise. To address these limitations, we propose RST-Guarder, an inference-time method that improves harmful-content detection for long-form inputs without additional data curation or model training. RST-Guarder first applies a RST parser to long-form inputs to get discourse-level semantic relations among segments, and subsequently performs hierarchical probabilistic inference to aggregate segment-level safety scores produced by pre-trained guardrail models. We evaluate RST-Guarder across multiple benchmarks and a diverse set of widely used guardrail models. Experimental results demonstrate that RST-Guarder consistently improves harmful-content detection on long-form inputs, while significantly reducing false positives that incorrectly classify benign content as harmful.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Xu Zhang, Hao Li Washington University, Saint Louis, Zhichao Lu City University of Hong Kong
Multimodal Large Language Models (MLLMs) achieve strong reasoning and perception capabilities but are increasingly vulnerable to jailbreak attacks. While existing work focuses on explicit attacks, where malicious content resides in a single modality, recent studies reveal implicit attacks, in which benign text and image inputs jointly express unsafe intent. Such joint-modal threats are difficult to detect and remain underexplored, largely due to the scarcity of high-quality implicit data. We propose ImpForge, an automated red-teaming pipeline that leverages reinforcement learning with tailored reward modules to generate diverse implicit samples across 14 domains. Building on this dataset, we further develop CrossGuard, an intent-aware safeguard providing robust and comprehensive defense against both explicit and implicit threats. Extensive experiments across safe and unsafe benchmarks, implicit and explicit attacks, and multiple out-of-domain settings demonstrate that CrossGuard significantly outperforms existing defenses, including advanced MLLMs and guardrails, achieving stronger security while maintaining high utility. This offers a balanced and practical solution for enhancing MLLM robustness against real-world multimodal threats. Our code is released https://github.com/ZhangXu0963/CrossGuard.
尚无 AI 解读，展示原文摘要
cs.CL · cs.CV
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Xiangxu Zhang Renmin University of China, Lei Li, Yanyun Zhou Renmin University of China, Xiao Zhou Renmin University of China, Yingying Zhang Tencent, Xian Wu Tencent
Medical diagnostics is a high-stakes and complex domain that is critical to patient care. However, current evaluations of large language models (LLMs) remain limited in capturing key challenges of clinical diagnostic scenarios. Most rely on benchmarks derived from public exams, raising contamination bias that can inflate performance, and they overlook the confounded nature of real consultations beyond textbook cases. Recent dynamic evaluations offer a promising alternative, but often remain insufficient for diagnosis-oriented benchmarking, with limited coverage of clinically grounded confounders and trustworthiness beyond accuracy. To address these gaps, we propose DyReMe, a dynamic benchmark for medical diagnostics that provides a controlled and scalable stress test of diagnostic robustness. Unlike static exam-style questions, DyReMe generates fresh, consultation-style cases that incorporate clinically grounded confounders, such as differential diagnoses and common misdiagnosis factors. It also varies expression styles to capture heterogeneous patient-style descriptions. Beyond accuracy, DyReMe evaluates LLMs on three additional clinically relevant dimensions: veracity, helpfulness, and consistency. Our experiments show that this dynamic approach yields more challenging assessments and exposes substantial weaknesses of stateof-the-art LLMs under clinically confounded diagnostic settings. These findings highlight the urgent need for evaluation frameworks that better assess trustworthy medical diagnostics under clinically grounded confounders.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Weixu Zhang McGill University and Mila - Quebec Artificial Intelligence Institute, Ye Yuan McGill University, McGill University and Mila - Quebec Artificial Intelligence Institute, Changjiang Han, Yuxing Tian Université de Montréal, Zipeng Sun, Linfeng Du, Jikun Kang SalesForce.com, Hong Kang 等 10 位作者
Large Language Models (LLMs) exhibit strong implicit personalization ability, yet most existing approaches treat this behavior as a black box, relying on prompt engineering or fine tuning on user data. In this work, we adopt a mechanistic interpretability perspective and hypothesize the existence of a sparse set of Preference Heads, attention heads that encode user specific stylistic and topical preferences and exert a causal influence on generation. We introduce Differential Preference Steering (DPS), a training free framework that (1) identifies Preference Heads through causal masking analysis and (2) leverages them for controllable and interpretable personalization at inference time. DPS computes a Preference Contribution Score (PCS) for each attention head, directly measuring its causal impact on user aligned outputs. During decoding, we contrast model predictions with and without Preference Heads, amplifying the difference between personalized and generic logits to selectively strengthen preference aligned continuations. Experiments on widely used personalization benchmarks across multiple LLMs demonstrate consistent gains in personalization fidelity while preserving content coherence and low computational overhead. Beyond empirical improvements, DPS provides a mechanistic explanation of where and how personalization emerges within transformer architectures.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Quyu Kong Alibaba Group, Xu Zhang Alibaba Group, Zhenyu Yang The Hong Kong University of Science and Technology, Nolan Gao, Chen Liu The University of Queensland, Panrong Tong Alibaba Cloud, Chenglin Cai Alibaba Group, Hanzhang Zhou 等 13 位作者
While AndroidWorld has become the dominant mobile-use benchmark due to its reproducible environment and deterministic evaluation, recent agents achieving over 90% success rates indicate saturation and motivate the need for greater challenge. In addition, its environment lacks key application categories, such as e-commerce and enterprise communication, and does not reflect realistic mobile-use scenarios characterized by vague user instructions and hybrid tool usage. We introduce MobileWorld, a substantially more challenging benchmark with 201 tasks across 20 applications that reflects real-world usage through long-horizon, cross-application workflows requiring nearly twice as many steps (27.8 vs. 14.3) and featuring significantly more multi-app tasks (62.2% vs. 9.5%) than AndroidWorld. MobileWorld balances production-grade utility and reproducible evaluation using open-source alternatives to industry standards (e.g., Mattermost for Slack), enabling full observability through source code modification and direct database access. Beyond standard GUI manipulation, MobileWorld introduces novel task categories including agent-user interaction and Model Context Protocol (MCP)-augmented tasks for evaluating agents in user-aware, hybrid-tool scenarios. We develop a planner-executor framework with extended action spaces supporting user interactions and MCP calls. Results show a sharp performance drop from AndroidWorld, with the best agentic framework and end-to-end model achieving 51.7% and 20.9% success rates, respectively, highlighting substantial room for future research.
尚无 AI 解读，展示原文摘要
cs.CL · cs.AI
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Zheyuan Zhang, Kaiwen Shi, Zhengqing Yuan, Zehong Wang, Tianyi Ma University of Notre Dame, Keerthiram Murugesan International Business Machines, Vincent Galassi, Chuxu Zhang University of Connecticut 等 9 位作者
Large language models (LLMs) and agent-based frameworks have advanced rapidly, enabling diverse applications. Yet, with the proliferation of models and agentic strategies, practitioners face substantial uncertainty in selecting the best configuration for a downstream task. Prior studies show that different agents and backbones exhibit complementary strengths, and that larger models are not always superior, underscoring the need for adaptive routing mechanisms. Existing approaches to agent routing, however, often emphasize cost efficiency while overlooking the fine-grained contextual and relational structure inherent in QA tasks. In this paper, we propose AgentRouter, a framework that formulates multi-agent QA as a knowledge-graph–guided routing problem supervised by empirical performance signals. Specifically, we convert QA instance into a heterogeneous knowledge graph that jointly encodes queries, contextual entities, and agents, and then train a heterogeneous graph neural network (GNN) to propagate information across node types and produce task-aware routing distributions over agents. By leveraging soft supervision and weighted aggregation of agent outputs, AgentRouter learns principled collaboration schemes that capture the complementary strengths of diverse agents. Extensive experiments demonstrate that our framework consistently outperforms single-agent and ensemble baselines, while generalizing across benchmarks and LLM backbones. These results highlight the effectiveness and robustness of graph-supervised multi-agent routing for question answering. Our code repo is available at https://anonymous.4open.science/r/AgentRouter.
尚无 AI 解读，展示原文摘要
cs.CL · cs.AI · cs.MA
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Xiaoying Le, Pengfei Qian, Yuanzhao Zhai, Xu Zhang National University of Defense Technology, Qian Liu National University of Defense Technology, Feng Dawei National University of Defense Technology, Bo Ding National University of Defense Technology
Scientific discovery evolution does not emerge in isolation but stems from the structural deepening and recombination of existing functionalities. However, current automated hypothesis generation methods, constrained by the statistical co-occurrence nature of Large Language Models (LLMs), lack perception of temporal causality and the “evolutionary patterns” inherent in scientific development. Consequently, they often yield superficial combinations that are logically infeasible. To address this, we propose EvoNarrator, a framework for hypothesis generation based on evolutionary narratives. We first extract structured P-M-L-F (Problem, Method, Limitation, Future Work) quadruples from citation networks. Subsequently, we introduce the SocketMatch mechanism, which eliminates logical disconnects between methods and problems by assessing their deep semantic compatibility. Finally, utilizing three macro patterns—Chain, Divergence, and Convergence—we constrain the generation process within historically logical derivation paths. Furthermore, double-blind expert reviews yielded an average score of 4.80/5.00 across novelty, feasibility, theoretical, and Logical. Additionally, hindcasting experiments validated its predictive foresight. Crucially, ablation studies indicate that integrating evolutionary patterns facilitates a paradigm shift from conservative incrementalism to theoretically grounded structural innovation. The code is available at https://github.com/xiyii-star/EvoNarrator.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Dongxu Zhang, Yiding Sun, Cheng Tan Shanghai Artificial Intelligence Laboratory and The Chinese University of Hong Kong, Wenbiao Yan, Ning Yang Institute of automation, Chinese academy of science, Chinese Academy of Sciences, Jihua Zhu Xi'an Jiaotong University, Haijun Zhang
While Chain-of-Thought (CoT) reasoning significantly enhances the performance of Multimodal Large Language Models (MLLMs), its autoregressive nature incurs prohibitive latency constraints. Current efforts to mitigate this via token compression often fail by blindly applying text-centric metrics to multimodal contexts. We identify a critical failure mode termed Visual Amnesia, where linguistically redundant tokens are erroneously pruned, leading to hallucinations. To address this, we introduce V-Skip that reformulates token pruning as a Visual-Anchored Information Bottleneck (VA-IB) optimization problem. V-Skip employs a dual-path gating mechanism that weighs token importance through both linguistic surprisal and cross-modal attention flow, effectively rescuing visually salient anchors. Extensive experiments on Qwen2-VL and Llama-3.2 families demonstrate that V-Skip achieves a speedup with negligible accuracy loss. Specifically, it preserves fine-grained visual details, outperforming other baselines over 30% on the DocVQA.
尚无 AI 解读，展示原文摘要
cs.CL · cs.CV
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
AI 速览：Plan-X 用可学习的多模态语言模型作为语义规划器，先生成时空语义 token 序列作为“语义草图”，再指导扩散 Transformer 生成与指令对齐的视频。
Diffusion Transformers have demonstrated remarkable capabilities in visual synthesis, yet they often struggle with high-level semantic reasoning and long-horizon planning. This limitation frequently leads to visual hallucinations and mis-alignments with user instructions, especially in scenarios involving complex scene understanding, human-object interactions, multi-stage actions, and in-context motion reasoning. To address these challenges, we propose Plan-X, a framework that explicitly enforces high-level semantic planning to instruct video generation process. At its core lies a Semantic Planner, a learnable multimodal language model that reasons over the user's intent from both text prompts and visual context, and autoregressively generates a sequence of text-grounded spatio-temporal semantic tokens. These semantic tokens, complementary to high-level text prompt guidance, serve as structured "semantic sketches" over time for the video diffusion model, which has its strength at synthesizing high-fidelity visual details. Plan-X effectively integrates the strength of language models in multimodal in-context reasoning and planning, together with the strength of diffusion models in photorealistic video synthesis. Extensive experiments demonstrate that our framework substantially reduces visual hallucinations and enables fine-grained, instruction-aligned video generation consistent with multimodal context.